Alcohol use and alcoholism are presently viewed as governed by antecedent factors that include childhood environment and genetically determined biological characteristics. Less attention has been paid, however, to how these antecedents influence later drinking. In addressing this issue, the psychosocial and biological literature both refer to acquired (learned) behavior, which is possibly influenced by differing reinforcement values of alcohol for different subsets of drinkers. It is, therefore, important to understand the precise nature of this learning and how it influences ongoing decisions about whether, and how much, to drink. Recent research on alcohol-related expectancy, considered in the context of cognitive (information processing) psychology, has opened a new research "window" on this learning. Exploration of the antecedent factors which influence expectancy development is just beginning, however, and there has been no exploration of the operation of expectancies immediately proximal to actual drinking opportunities. The present project will comprehensively investigate whether alcohol expectancies may be one (mediational) mechanism by which psychosocial and biological antecedents these influence later drinking behavior. Three research strategies for investigating mediational relationships will be used. A prospective design, using covariance structure modeling (LISREL) techniques, will assess the capacity of alcohol expectancy to serve as a psychological mediator. Since expectancies are theorized to influence drinking at a point proximal to a drinking opportunity, a study of the feasibility of monitoring expectancies in the natural environment, proximal to actual drinking occasions, will be also undertaken. Finally, since the most powerful test of mediation is effective experimental manipulation, an extensive experimental challenge of alcohol expectancies will also be conducted. Specifically, this research program will: assess whether antecedent variables which have been shown to influence later drinking can be usefully understood as carrying over their influence by psychological/information processing mechanisms; increase our understanding of which antecedent variables most influence the psychological mechanisms; assess expectancy operation proximal to drinking decisions; test the possibility of actually manipulating drinking via challenging expectancies; and refine expectancy challenge procedures for application to primary and secondary prevention.